linux-kernel - Re: [PATCH 3/3] x86/MCE/AMD: Get address from already initialized block

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20180517134415.GC27738@pd.tnic>
Date:   Thu, 17 May 2018 15:44:15 +0200
From:   Borislav Petkov <bp@...e.de>
To:     "Ghannam, Yazen" <Yazen.Ghannam@....com>
Cc:     Johannes Hirte <johannes.hirte@...enkhaos.de>,
        "linux-edac@...r.kernel.org" <linux-edac@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "tony.luck@...el.com" <tony.luck@...el.com>,
        "x86@...nel.org" <x86@...nel.org>
Subject: Re: [PATCH 3/3] x86/MCE/AMD: Get address from already initialized
 block

On Thu, May 17, 2018 at 01:04:19PM +0000, Ghannam, Yazen wrote:
> Yes, you're right. I thought using the existing data structures would work, but it
> seems I messed up the implementation.

Not only that - your idea wouldn't fly because the per-CPU stuff you
were using gets torn down when the CPU goes offline so you can't use
them on resume because they're not there yet.

> Banks 15 and 16 should have an address for block 1 also. Do you have PFEH
> enabled on your system? That would cause MISC0 to be RAZ so we won't
> get the MISC1 address. I'll try it myself also and let you know.

I check PFEH is enabled how?

> I think this good for now. We'll probably need to change it in the
> future, but maybe we can clean up all the thresholding blocks code and
> make it simpler when we do change it.

Ok.

> This hunk could go above the !block. Though maybe the macro is lighter than the
> array lookup. It'll work either way, but I'm just thinking out loud.

Yeah, it doesn't matter in that case.

> Since we're caching the values during init, we can drop all the
> *_on_cpu() calls. What do you think?

Well, if they're all the same on all CPUs, sure. That's your call.

-- 
Regards/Gruss,
    Boris.

SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg)
--